互相关函数 Resources

Showing items tagged with "互相关函数"

This project implements speech recognition for passwords composed of digits 0-9 by calculating the cross-correlation function between two signals and their corresponding variance, using maximum variance as the discrimination threshold. Signal preprocessing is essential and includes FIR filter-based pre-emphasis and endpoint detection using short-term average energy and zero-crossing rate methods to extract useful signals. These preprocessing steps maximize recognition accuracy. The cross-correlation function, which quantifies signal similarity, helps differentiate test signals from template signals. In MATLAB implementation, key functions like xcorr() for cross-correlation calculation and var() for variance computation are utilized, while preprocessing involves designing FIR filters with fir1() and implementing energy/zcr-based endpoint detection algorithms.

MATLAB 257 views Tagged